This study uses retrospective illness histories to investigate whether children’s health deteriorates after parental separation. Separation is associated with illness in a multivariate cross-sectional analysis as well as in an analysis of a sample of disrupted children only, in which illness rates before and after separation are compared. Three explanations are hypothesized: (1) divorce reduces the resources available to children, (2) the stress of divorce depletes children’s health, and (3) frailer children are selected into divorce. The first hypothesis has stronger support than the second, but the data are too poor for a rigorous test of either. The selection hypothesis is not supported.